Americans sour on AI while Google, Anthropic, and robots make moves
Only 16% of Americans think AI is a net positive — but that's not stopping Google, Khosla, and Canadian pension funds from going all in.
Subscribe free →Only 1 in 6 Americans Thinks AI Is Good for Society
Pew Research dropped a brutal number on the industry today: just 16% of Americans believe AI will have a positive impact on society. That's a stark disconnect from the billions being poured into the space by Wall Street and venture capital. The gap between investor enthusiasm and public trust has never looked wider.
Why it matters: AI companies need social license to operate at scale — and they're burning through it fast without much effort to earn it back.
Anthropic's Fight With the White House Is Actually Boosting Its Business
Anthropic has been trading punches with the Trump administration over its AI storyworld project Fable, and according to spending data from Ramp, enterprise customers are responding by opening their wallets wider. The government beef appears to be lending Anthropic a contrarian credibility with business buyers who see Claude as the anti-establishment enterprise option.
Why it matters: Political friction is becoming a marketing asset — and Anthropic is proof that picking the right fight can drive B2B revenue.
Google Kills the Dumb Smart Speaker With a $99 Gemini-Powered Home Device
Google's new Home Speaker ditches the old "Hey Google, set a timer" rigidity for full conversational Gemini interactions. At $99.99, it's a direct shot at the Echo lineup, and it's the clearest signal yet that Google is betting the home audio category can be rebooted by replacing command parsing with actual language understanding.
Why it matters: If Gemini delivers on natural conversation in the living room, it resets expectations for every smart home device — and puts Amazon's Alexa upgrade timeline under serious pressure.
The Unglamorous Data Crisis Blocking the Robot Revolution
Physical AI has a dirty secret: getting robots to work in the real world requires enormous amounts of tedious, human-generated training data that no one wants to collect. AI labs are already paying companies like XDOF to do the grunt work — hours of humans demonstrating tasks in warehouses and kitchens so robots can learn from the footage. Without solving this data pipeline problem, humanoid robotics won't replicate what LLMs pulled off with text.
Why it matters: Robot data collection is the new data labeling industry — and whoever cracks the pipeline at scale will have a massive moat in physical AI.
Khosla Bets $27M That AI Needs Mathematical Proof, Not Just Vibes
Pramaana Labs just closed a $27M seed round from Khosla Ventures to bring formal verification to AI outputs — essentially applying mathematical proof techniques to guarantee correctness in high-stakes domains like law, drug discovery, and tax prep. The idea is that for verticals where a hallucination could cost millions or lives, "pretty good" accuracy isn't good enough.
Why it matters: Formal verification could be the unlock that gets AI into regulated industries that have stayed on the sidelines precisely because they can't tolerate unpredictable errors.
Canadian Pension Money Is Flowing Into India's Data Center Boom
A major Canadian pension fund is acquiring an 8.2% stake in CtrlS, which runs 15+ data centers across India. This isn't a tech company bet — it's institutional infrastructure capital treating AI compute capacity like toll roads and pipelines: boring, essential, and lucrative over decades.
Why it matters: When pension funds start buying AI infrastructure stakes, it signals the market views data center demand as a long-term structural trend, not a bubble.
Quick Hits
- →Android 17 is out with expanded Gemini features baked into Pixel devices via a same-day Pixel Drop — Google is tightening the AI-OS loop fast. TechCrunch
- →Pinterest launched **Ask Pinterest**, a standalone conversational AI shopping app — yet another legacy platform trying to bolt a chat interface onto its discovery engine. TechCrunch
- →DeepL acquired live-event audio startup Mixhalo, opening a San Francisco office as it pushes real-time spoken translation into stadiums and conferences. TechCrunch
- →Qualcomm says it's working with partners on 40+ AI wearable devices — earbuds with cameras, jewelry, pins — as it positions itself as the chip inside whatever kills the smartphone. TechCrunch
- →Mobileye is launching its own US robotaxi service, putting it in direct competition with the automakers it sells its self-driving tech to — an awkward but aggressive pivot. TechCrunch
AI TLDR